Reading Paths

Abstract

Seven ordered routes through the encyclopedia for readers who want a story rather than an index. Each path is a sequence of existing articles, ten to twenty long, with one line on why the next article follows. The paths overlap on purpose: the Morris worm is a security story, a hacking story, and a network story. For a year-by-year route instead, start at the timeline; for the thematic map, [[index|the index]].

How to use these

Read in order. The first article of each path is chosen to need no prerequisites; the later ones assume the earlier. Side trips are listed at the end of each path, not inside it.

Security in Twelve Articles

From the mathematics that made secrets keepable to the discovery that the hardware itself leaks.

Cryptography: The Secret Science — the overview, from Caesar to the quantum threat; sets the vocabulary.
Whitfield Diffie and Martin Hellman — 1976: two strangers agree a secret over a public wire.
Rivest, Shamir, Adleman, and RSA — 1977: the trapdoor that turned the idea into a padlock.
Robert Tappan Morris — 1988: the worm that ended the trusting network.
The KGB Hack — 1986: a 75-cent accounting error uncovers state espionage.
Phil Zimmermann and PGP — 1991: strong encryption for anyone, and a criminal investigation.
The Crypto Wars of the 1990s — Clipper, “code is speech”, and Deep Crack.
Secure by Design — the defenders’ history: Saltzer and Schroeder, the buffer overflow, Gates’s memo.
The Vulnerability Disclosure Debate — what to do when you find the bug.
The History of Cyberwar — Stuxnet: code that breaks centrifuges.
Edward Snowden and the NSA — 2013: the scale of collection, and the encryption it provoked.
Spectre, Meltdown, and Hardware Vulnerabilities — 2018: the processor’s own optimizations as attack surface.

Side trips: Paul Karger (the 1974 report that predicted it all), Kevin Mitnick, The Ransomware Epidemic, Bruce Schneier, The History of Passwords and Authentication.

AI from the Perceptron to Claude

Two ideas, symbols and statistics, taking turns to be the future for seventy years.

The Rise of Artificial Intelligence — the overview: Dartmouth, two winters, the statistical turn.
Frank Rosenblatt and the Perceptron — 1957: a machine that learns, and the book that buried it.
Newell and Simon — 1956: the symbolic alternative, the Logic Theorist.
Marvin Minsky and the MIT AI Lab — the man who co-founded the field and argued neural networks were a dead end.
Joseph Weizenbaum and ELIZA — 1966: the first chatbot, and its inventor’s alarm.
Expert Systems and the First AI Winter — the 1980s boom in hand-coded knowledge and its 1987 collapse.
John Hopfield and Neural Networks — 1982: memory as energy minimization; the physicists arrive.
Geoffrey Hinton and Deep Learning — thirty years of unfashionable backpropagation.
Schmidhuber, Hochreiter, and LSTM — 1997: the network that could remember.
Fei-Fei Li and ImageNet — the dataset that made the proof possible.
Alex Krizhevsky — 2012: AlexNet on two gaming cards in a bedroom.
The GPU Revolution — why the hardware was ready.
Barto and Sutton — reinforcement learning, the third paradigm.
DeepMind and AlphaGo — 2016: Move 37.
The Transformer Architecture — 2017: attention is all you need.
Ilya Sutskever and the GPT Series — scaling as a research program.
Sam Altman and OpenAI — ChatGPT, 100 million users in two months.
The LLM Race — the competitive cascade that followed.
Anthropic and Claude — the safety lab that races.
AI Ethics and Algorithmic Bias — what the systems do to the people they are used on.

Side trips: Alan Turing for the 1950 paper that started it, Douglas Lenat and Cyc for the symbolic road taken to its end, Judea Pearl for the critique from causality, AlphaFold for AI as science.

The Germany Path

A country that built the first program-controlled computer, coined Informatik, wrote the first data-protection law, and left the consumer internet to others.

Konrad Zuse and the Z3 — 1941: the first working Turing-complete computer, built in a Berlin workshop.
Helmut Schreyer — the friend who wanted to make it electronic, and was told the war would be over first.
Germany’s Computing Pioneers — Leibniz to Petri nets: the survey.
GI: Gesellschaft für Informatik — 1969: the discipline declares itself.
German Chapter of the ACM — the older, smaller, international sibling.
Heinz Nixdorf — Mittlere Datentechnik for the Mittelstand; death on the CeBIT dance floor.
Dead End: Buying a Computer Industry — Unidata, the European IBM that France walked out of.
Computing in East Germany — cloning under embargo, the megabit chip, the collapse of 1990.
SAP — five IBM engineers in Weinheim build Europe’s largest software company.
Dead End: Bildschirmtext — the Bundespost’s pay-per-page online service.
Wau Holland — the Chaos Computer Club’s civic hacker ethic, and the Btx hack.
The KGB Hack — the CCC milieu’s darkest chapter.
CeBIT — Hannover’s 830,000-visitor fair and why it died.
Günter von Gravenreuth — the Abmahnung lawyer who shaped a generation’s view of copyright.
Karlheinz Brandenburg and MP3 — Erlangen’s codec conquers the world through a stolen encoder.
Hans Uszkoreit and German Language Technology — DFKI, Verbmobil, and the son who co-wrote the Transformer paper.
Germany’s IT Industry — the pattern: enormous capability, no consumer platform.

Side trips: Heinz Zemanek (Vienna), Computer Art and Generative Art (Stuttgart 1965), Gustav Tauschek, Fuzzy Logic for a contrast in how ideas travel.

The Network, from Telegraph to Browser

Data Communication Prehistory — the telegraph invents encoding, multiplexing, and store-and-forward.
Paul Baran and Packet Switching — the survivable mesh that was never built.
J.C.R. Licklider — the funding vision.
Larry Roberts and ARPANET — the engineer who built it.
ARPANET: Building the Network — “LO”, 29 October 1969.
Louis Pouzin and the Datagram — the French network whose argument won inside TCP/IP.
Vint Cerf and Bob Kahn — TCP/IP and Flag Day 1983.
Hubert Zimmermann and the Seven Layers — the OSI model that named everything and shipped nothing.
Early Networking Failures — Token Ring, ATM, OSI: what lost.
Bob Metcalfe and Ethernet — the local network that never got replaced.
The Domain Name System — names instead of numbers.
Jon Postel — the man who kept the lists.
Tim Berners-Lee and the World Wide Web — “Vague but exciting”.
Dead End: Gopher — the rival that a licensing decision killed.
Mosaic — inline images, 1993.
The Browser Wars — Netscape, Microsoft, Firefox, Chrome.

The Personal Computer in Twelve Articles

The Altair 8800 — January 1975: the kit on the magazine cover.
The Open Letter to Hobbyists — 1976: software as property.
Steve Wozniak — the Apple I and II, designed alone.
The Home Computer Boom — the 1977 trinity and the 1983 price war.
Dan Bricklin and VisiCalc — the program that sold the hardware.
Gary Kildall and the OS IBM Didn’t Take — CP/M and the missed meeting.
The IBM PC — August 1981: open architecture, borrowed operating system.
Compaq and the BIOS Clean Room — how the clone became the standard.
The Lotus 1-2-3 Story — the PC becomes a business necessity.
The Macintosh — 1984: the graphical machine an individual could buy.
The Windows Story — from derided shell to seven in ten desktops.
The Commodity PC — where the margin went.

Then the spine that puts them in order: The Personal Computing Explosion.

Theory, from Hilbert to P vs NP

David Hilbert and the Entscheidungsproblem — 1928: the question.
Gödel and the Incompleteness Theorems — 1931: the first no.
Alonzo Church and Lambda Calculus — 1936: the second no, by functions.
Alan Turing and the Enigma — 1936: the third no, by machines.
Automata Theory and Computability — the ladder of machines.
The Chomsky Hierarchy — grammars matched to machines.
Hartmanis and Stearns — 1965: measuring the cost of computation.
Jack Edmonds and the Good Algorithm — polynomial versus exponential, named.
Cook and Levin — 1971: NP-completeness, twice.
Richard Karp — 1972: the 21 problems.
P vs. NP and Complexity Theory — the open question.
SAT Solvers — the hard problem solved daily anyway.
The Busy Beaver Problem — uncomputability as a countable sequence.
Kolmogorov Complexity — information as the shortest program.

Dead Ends, a Starter Kit

Ten failures chosen because each fails for a different reason.

Dead End: Gopher — a licensing announcement.
Dead End: OS/2 — technical superiority in a platform war.
Dead End: BeOS — one negotiation.
Dead End: General Magic — fifteen years early.
Dead End: Symbian — won every battle, lost to a redefinition.
Dead End: Iridium — the venture died, the machine in orbit did not.
Dead End: The Transputer — twenty years early for multicore.
Dead End: Google Reader — killed while working, and what that taught users.
Dead End: Buying a Computer Industry — four programmes, one method.
Route 128 — a dead end that is a place.

The full list is under Part XII of [[index|the index]].

📚 Sources

This page makes no claims of its own; every fact and date belongs to the linked article and is sourced there.